OCT Camera Position Affects Apparent Eye Morphometry in Tree Shrews (Tupaia belangeri)

Abstract

Purpose : Optical coherence tomography (OCT) machines are calibrated for human eyes but are used in research to image small animals. Morphologic remodeling is expected in some animal models of disease, such as glaucoma or myopia. Serial in-vivo OCT imaging is used to quantify these longitudinal morphologic changes. However, change in camera position can create optical distortions and false positive morphometric changes. This study quantifies apparent morphometric changes in tree shrews evoked by OCT camera movement along the visual axis.

Methods : Similar to Garrison et. al., x-axis scaling (along a b-scan) was calibrated by injecting 100µm glass beads onto the retina (4 eyes of 2 animals). Factory z-axis scaling (along the visual axis) was used. 14 eyes of 9 different tree shrews were imaged (Spectralis OCT2; Heidelberg Engineering, Inc). Baseline horizontal and vertical scans (30^o, ART 15, EDI) through the optic nerve head (ONH) were obtained and repeated without follow-up mode after moving the camera anterior and posterior 5mm. B-scans were delineated with Multiview software. For each b-scan, 3 measures were obtained: 1) Bruch’s membrane opening (BMO), 2) retinal thickness (RT) above BMO, and 3) RPE position 500µm lateral to BMO relative to a line through the BMO points.

Results : Camera movement anterior and posterior significantly affected BMO compared to baseline (+44±5µm p<0.001 & -47±5µm p<0.001, respectively). RT above BMO was different with anterior movement (+9±3µm p=0.02) not posterior (-3±3µm p=1.0). Greatest effects were seen as apparent sclero-retinal bowing (Figure1). Position of the RPE relative to the BMO line increases with anterior movement (+55±4µm p<0.001) and decreases with posterior movement (-78±4µm p<0.001).

Conclusions : Follow-up mode on the Spectralis likely uses image based matching algorithms that can mask the true disease-related morphologic changes. However, this study provides evidence that camera position changes cause significant non-linear false positive morphometric changes in the x-axis when follow-up mode is not used. Accurate morphometric measures in longitudinal studies likely require use of a species specific optical compensation algorithm.

This abstract was presented at the 2019 ARVO Annual Meeting, held in Vancouver, Canada, April 28 - May 2, 2019.

 

B-scans through the tree shrew ONH showing the effect of the OCT camera position on morphology. Compared to baseline (B), moving the camera either (A) 5 mm anterior or (C) 5mm posterior creates convex or concave distortions, respectively.